|
violet Posts: 15
4/30/2012
|
***Resolved on 5/3/12***
Hello,
I am using StressStimulus (Version 1.7.4497 beta) to perform load testing. I am using the new feature "request once" which defines the initial actions in the test scenario. I am also using a parameterized value. what I do is I assign this variable to all of my requests. my tests case contains 9 urls, out of which only the last 2 urls need to be repeated over and over (i .e 7 urls need to be requested once). I have noticed in the second round of iterations, the variable is not assigned to the corresponding value in my last two urls (value request has not been paramatized correctly) and it defaults back to the recorded value.
Is it me doing sth wrong or is it a bug?
Many thanks,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
4/30/2012
|
Hi Violet,
I think I know what is happening here. Currently, the life-span of an extractor is one iteration. Once the iteration is complete, the extractor’s value becomes undefined. Before the "request once" feature was implemented, it was OK because on the subsequent iterations the new extractor value is re-created.
With the new "request once" feature, an extractor from the request, which runs only on the 1-st iteration, becomes undefined on the subsequent iterations. Is a result, the parameterization will not work.
To address this issue we need to expand the extractor life-span from iteration to the entire test.
Does this sound like a correct guess?
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
4/30/2012
|
Hi there,
yes, that is exactly what is happening .
will it take long to be fixed?
Many thanks,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
4/30/2012
|
Thanks for letting me know. The fix should be ready later today. I will let you know as soon as it is released.
Cheers,
-Vadim
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
4/30/2012
|
We just released the new build with the fix, available at http://www.stresstimulus.com/default.aspx#dwnld Please let me know if it fixed the parameterization.
Thanks,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
4/30/2012
|
Hi Vadim,
Thanks for the quick fix. That issue is now solved.However, I noticed another issue which might not be related to "request once" feature and that is :
As i said earlier, I am using a parameterized variable in the URL and query section of my requests. lets say I am parameterizing the connection handler "c" and assigning it to the variable "conn". The "c" variable is also being used in the body section of my Post requests. I saw that its value is {{Auto-Correlated}} by default.
Running the test, I noticed that the value of "c"in the body section is the same as the recorded value which should not be so and is wrong. I thought that {{Auto-Correlated}} would some how correlate the value of "c" to the updated value (current connection handler), but it seemed i was wrong.
when I assigned the value "conn" into "c" in the body section of the post, it all went wrong and even the "conn" variable were not assigned correctly to the URL queries.
Any possible explanation? and what does {{Auto-Correlated}} do exactly?
Thanks again for your prompt reply.
Best,
|
|
|
0
link
|
|
violet Posts: 15
4/30/2012
|
Hi again,
Having downloaded the newest version(Version 1.7.4503 beta) to apply the fix, I tested some scenarios without checking the "Request once" option which is iterating the test cases over and over. there were some cases that the parametized variable was not consistent in the life span of the iterate. And some iterations were missing the parametized variable (undefined). I think it is as the side-effect of fixing that bug, is not it so?
Best,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
4/30/2012
|
I can check the specific issues that you described if you upload your test to our support repository. To do so, click "upload support case" in the StresStimulus menu -> Support Options. Please describe the issue and indicate the requests that are affected.
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
5/1/2012
|
Hello,
The IT team here are preparing a public face server to enable you to use my test case. Meanwhile, I also noticed that using the"request once" feature, only the first virtual user gets the value from the extraction and the parameterized variable is not set for the following virtual users (2, 3..). You may be able to reproduce that case yourself.
Best,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/1/2012
|
Hi,
Thanks for letting me know. Once you upload the test, I will look into this behavior as well.
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
5/1/2012
|
Hello, I uploaded a few cases which were not running as expected. Unfortunately or maybe fortunately I could not reproduce the last case!!!. Could you please let me know when you are finished so the IT team can make the server internal once more.
Many thanks,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/1/2012
|
I looked at your test case is and checked the two issues that you've described.
1. Parameterization of the field "C". I saw this issue. Here’s how to fix it. In the request #8, the field "C" is set to use auto-correlation. It works on the 1-st iteration by carrying over the value from the response #7 to the request #8. However on the 2-nd iteration, requests 1-7 are not issued, because you've set them as "run once", and thus the auto-correlation breaks. To fix it, replace the auto-correlation with the existing extractor “conn” in the request #8.
In the next build will allow the auto-correlated values to be carried over from the last response of the previous iteration to first requests of the subsequent iteration. As a result this setting with the optional.
2. The parameterization for the subsequent VUs. I cannot confirm that there is an issue with VUs 2,3,.. Once you use the setting described above, you should have no issues with the subsequent VUs.
Let me know if this was helpful.
Cheers,
-Vadim
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/1/2012
|
My response above is about your 1-st Test upload. I believe I've answered your questions. Do you need me to look at the 2-nd upload as well? If not and you are satisfied with the response, please go ahead and remove the access to your website.
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
5/1/2012
|
Thanks, your recommendation worked fine for test case 1.
about the second uploaded test, I appreciate if you have a look at it.It still does not look good to me.
Many thanks,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/1/2012
|
Sure. I will take a look. What was the issue with the 2-nd Test?
|
|
|
0
link
|
|
violet Posts: 15
5/1/2012
|
The requests were not sent in order in some iterations and in some cases some requests were sent twice (while it has not been the case in the test case ) hence resulting in new value for "conn" .
plz let me know if you have more questions.
Best,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/2/2012
|
I saw the issue. For whatever reason, your 2-nd test was saved incorrectly.
We just released a new build that double-checks that the tests are saved correctly.
Please install the new version and let me know if everything works well now.
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
5/2/2012
|
Hello,
I uploaded the 2-nd test once more.
please let me know if you have any questions regarding the test.
Best,
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
5/2/2012
|
Hi,
I checked the test that you've uploaded today, and now I can tell that the test was saved and opened correctly. However, when the test runs, your application returns the 3-rd request with the error message:
"There are no Logic Servers available to complete your request. Please try again later\n\nProblem Area:\nError Source - Business Logic Router\n\nUnable to locate a Business Logic Server which can invoke 'Fdportal.FDTek.FDPortal.UserControl.Request'. Ensure that the Business Logic Server which supports this event is available.\n\nClick OK for more information"
You need to troubleshoot it.
Hint: compare the recorded and the replayed requests 1-3 to see if all parameters are sent the way your application expects.
Cheers,
-Vadim
|
|
|
0
link
|
|
violet Posts: 15
5/2/2012
|
yes, sorry that "error" is solved now.Plz ignore that error as the server is up and running again. That was caused because i was doing some tests. Apparently it was not a good idea we both do the test at the same time. I am halting my tests to avoid interfering with yours. hopefully this time, things became clearer on the mentioned issue.
plz let me know if you had any questions.
Best,
|
|
|
0
link
|